The first commercially successful device using speech recognition was sold in 1922. A toy called "Radio Rex" consisted of a celluloid dog with an iron base. The dog sat in his doghouse held by an electromagnet which pressed against a spring. The current which energized the magnet flowed through a metal bar that formed a bridge with two supporting members. When this bridge was exposed to acoustic energy at 500 hertz, the current was interrupted and the dog sprang from his house. The vowel in Rex when spoken by most people creates a tone around 500 hertz. However, the toy did not truly recognize speech; it was unable to distinguish different commands.
The first efforts at attempting machine translation of speech came in the late 1940's as the US government was trying to transcribe and translate Russian documents. (The agency responsible for the research later came to be known as the Defense Advanced Research Project Agency - DARPA - which also presided over the invention of TCP/IP, which brought us the Internet.)
The earliest development of machine translation can be traced to conversations and correspondence between Andrew D. Booth, a British crystallographer, and Warren Weaver of the Rockefeller Foundation in 1947, and more specifically to a memorandum written by Weaver in 1949 to the Rockerfeller Foundation which included the following two sentences.
"I have a text in front of me which is written in Russian but I am going to pretend that it is really written in English and that it has been coded in some strange symbols. All I need to do is strip off the code in order to retrieve the information contained in the text."
True speech recognition was not available until 1952; Bell labs created a machine system that could distinguish the spoken numerals 0-9. By 1960, this system was upgraded to recognize - 60 words.
No comments:
Post a Comment